social capital and diversity final for PS that diversity, and racial heterogeneity in particular, is...

I would like to thank Geoffrey Evans, Christopher Garner, Sharon Gilad, Oliver Heath,

Michelle Jackson and Laura Stoker for their helpful comments and suggestions, and Giselle

Baker and Patten Smith from BMRB for merging the datasets. The paper has also benefited

from the comments by the three Reviewers and the Editor of Political Studies. The paper was

written while the author was a Postdoctoral Prize Research Fellow at Nuffield College, Oxford.

I would like to thank Nuffield College for their generous academic support.

Does Diversity Erode Social Cohesion? Social Capital and Race in British Neighbourhoods

Natalia Letki

Department of Political Science Collegium Civitas

Palace of Culture and Science, 12th floor 1 Plac Defilad

00-901 Warsaw, Poland [email protected]

tel. (++48 22) 656 71 93 fax (++48 22) 656 71 75

Abstract

The debate on causes and consequences of social capital has been recently

complemented with an investigation into factors that erode it. Various scholars

concluded that diversity, and racial heterogeneity in particular, is damaging for the

sense of community, interpersonal trust and formal and informal interactions.

However, most of this research does not adequately account for the negative effect

of a community’s low socio-economic status on neighbourhood interactions and

attitudes. This paper is the first to date empirical examination of the impact of

racial context on various dimensions of social capital in British neighbourhoods.

Findings show that the low neighbourhood status is the key element undermining

all dimensions of social capital, while eroding effect of racial diversity is limited.

forthcoming in the Political Studies

There is evidence that the more diverse an area is in racial terms, the less likely its residents

are to feel that they trust each other. This is an important argument and it is important that

we examine it.

From a speech by David Blunkett MP, Home Secretary,

to the Institute of Public Policy Research, 7th July 2004

Introduction

Changing patterns of immigration, perceptions of the increase in the numbers of

refugees and asylum seekers and social unrests in Northern England in 2001 (in the towns of

Oldham and Burnley and the city of Bradford) have brought the relationship between

community cohesion and ethnic diversity to the forefront of public and political debate. Fears

about the overwhelming and negative effect of diversity on social cohesion and national

identity have been expressed by journalists and policy makers alike. For example, David

Goodhart, the editor of The Prospect, targeted his February 2004 article at the detrimental

impact of ethnic diversity in modern Britain on the sense of community and solidarity among

citizens, and on the viability of the British welfare state (Goodhart 2004), while Trevor Philips,

Chair of the Commission for Racial Equality, reiterated this argument by stressing the need for

strengthening common values and “core Britishness” (Baldwin and Rozenberg 2004).1

However, while this public debate seems highly politicized and emotion-based, it is

important to look at the link between community cohesion and diversity in a more systematic

way. This paper examines the impact of racial diversity on one particular dimension of social

cohesion: social capital. Social cohesion is usually defined in reference to common aims and

objectives, social order, social solidarity, and the sense of place attachment (Forrest and

Kearns 2001). Social capital, i.e. “features of social life – networks, norms and trust – that

enable participants to act together more effectively to pursue shared objectives”, constitutes

therefore its key dimension (Putnam 1996: 56, see also Forrest and Kearns 2001; McGhee

2003). Although social capital is a desired resource that both individuals and communities can

use for good ends (Fukuyama 1995; Putnam 1993a), it seems that for social capital to emerge,

a high degree of homogeneity is required: empirical evidence suggests that communities with

high levels of racial and cultural diversity have lower levels of interpersonal trust and formal

and informal networks (Alesina and Ferrara 2000, 2002; Costa and Kahn 2003).

1 For the discussion of Labour Government’s attempts at “re-packaging and re-branding of

the 'British people'” and its “‘super market sweep’ tendencies in relation to social theory; that

is, the highly selective ‘smash and grab’ deployment and understanding of concepts and social

theory in New Labour’s legacy of public policy making”, see McGhee 2003: 377-380.

3

These observations prompted policy makers to abandon the dominant approach of

multiculturalism in favour of what some call “a return to assimilation” (Cheong et al. 2005: 2).

This shift has been reflected in major government policies as well as in public opinion, with

the current dominant political view stressing that “strength in diversity” should be achieved

through promoting shared values and creation of “unity from diversity” (Cantle 2001; Denham

2001; McGhee 2003). As a result, the political and social climate emerged, “in which asylum

seekers, migrants and refugees are demonized as undesirable, undeserving, and overwhelming”

(Ouseley 2004: 9).

Although the debate about the impact of ethnic fractionalization on social cohesion and

social ties seems fairly developed and its impact can be noticed far outside of academia, the

empirical evidence for the eroding effect of ethnic heterogeneity is mixed, and largely confined

to American examples. Recently, researchers investigating the impact of neighbourhood racial

and ethnic composition on individuals’ attitudes and behaviour started complementing the race

or ethnicity-related predictors with the socio-economic characteristics of an area. They present

evidence that socio-economic status of a neighbourhood affects interactions with, and attitudes

towards, fellow neighbours. Disorder and poverty negatively influence individuals’ ability and

willingness to engage in social activities with neighbours, they amplify the sense of

powerlessness and mistrust, and amplify inter-group prejudice and competition (Li et al. 2005;

Marschall and Stolle 2004; Oliver and Mandelberg 2000; Oliver and Wong 2003; Ross et al.

2001; Sampson et al. 1997). It is also an empirical fact that poverty and disorder tend to be

highly correlated with racial diversity (Sampson and Groves 1989; Sampson et al. 1997). In

this paper I argue that to properly assess the effect of racial diversity on social capital, such

strong correlation between diversity and low neighbourhood status needs to be taken into

account. Therefore, in explaining levels of interaction within community and attitudes that

result from them, deprivation should be treated as an equally important neighbourhood

characteristic as racial diversity.

Socio-economic inequalities, originally at the centre of investigation in political science

and sociology, have been denied importance by many researchers since 1980s. The “age of

'postisms'— postmodernism, postindustrialism, poststructuralism, postmarxism, postfeminism,

postmaterialism” (Evans 1999: 1) has shifted attention to mechanisms focusing on values and

new social divisions. The traditional cleavages of class and social status have been thought to

be replaced by “new” divisions structured around gender, ethnicity or ecology (Inglehart 1990,

Pakulski and Waters 1995). Although some researchers resisted and criticised this trend

(Devine et al. 2003; Erickson and Goldthorpe 1992), the intellectual fashion of denying

“traditional” social structure importance in explaining political and civic attitudes and

4

behaviour has swept through academia and beyond. However, recent studies show that such a

change of perspective was not fully justified empirically, and that social inequalities and social

deprivation still offer powerful explanations of political attitudes and behaviour (Evans 1999).

This present paper will contribute to this trend reversal by demonstrating that despite the

claims that cultural differences are the key issue undermining social cohesion, socio-economic

structure is far more important.

The paper presents an important contribution to the current debate on social cohesion

and social capital for three main reasons: firstly, it uses a complex and multi-faceted measure

of neighbourhood social capital; secondly, it tests the impact of respondents’ actual immediate

community on their attitudes; and thirdly, it applies a methodology that overcomes the

limitations of previous studies. Existing research on social capital or social trust usually

separates the individual and community level of analysis or refers to individual-level

explanations to interpret the aggregate-level findings. Also accuracy of the empirical

operationalisation of the relationships between dimensions of social capital is frequently

compromised. This paper applies a multi-level structural equation model, which accounts for

the hierarchical structure and path relationships of the theoretical propositions, thus

overcoming conceptual and methodological problems of earlier investigations. Moreover, it is

the first study analyzing the effects of racial diversity on social capital in the context of

neighbourhood in Great Britain: existing literature on the topic has been dominated by the

evidence from the United States.2 The main question of this paper is therefore: does racial

diversity erode social capital in Britain?

The paper falls into four main parts: First, I will outline the general theoretical

background to studying the relationship between community’s social capital and racial

diversity. Second, I will introduce data and indicators. After that I will explain the method used

to explore the question about the link between social capital and racial diversity. Fourthly, I

will present the analyses and discuss their results.

The results show that when the effects of diversity and neighbourhood deprivation on

social capital are modelled simultaneously, and the relationship between neighbourhood status

and racial diversity is accounted for, diversity has a negative effect on only one – attitudinal -

dimension of social capital. At the same time, the low socio-economic status of a

neighbourhood is the main factor undermining all types of interactions and positive attitudes

among neighbours. Findings presented in this paper show that the British government’s efforts

2 The two main studies of social capital in Britain focus on aggregate patterns over time or

individual-level multiple indicator assessment, but they do not investigate social capital in the

context of local community’s ethnic diversity. See Hall 1999 and Li et al. 2005.

5

to de-emphasize socio-economic deprivation and focus on stimulating inter-community

relationships and creating “unity from diversity” have been rather misplaced, as deprivation is

the major factor eroding community relations and negatively tinting diversity.

Social capital: What is it and why it matters?

Social capital is associated with “people's sense of community, their sense of belonging

to a neighbourhood, caring about the people who live there, and believing that people who live

there care about them” (Portney and Berry 2001: 71). Positive attitudes towards and beliefs

about one’s neighbours contribute to cohesion within local community, and thus to residents’

willingness to participate in local affairs and to cooperate in everyday matters. As a result, life

in communities with high levels of social capital – so called “civic communities” – is good:

“the presence of social capital - individuals connected to one another through trusting networks

and common values - allows for the enforcement of positive standards for youths and offers them

access to mentors, role models, educational sponsors, and job contacts outside the neighbourhood.

Social networks may also provide emotional and financial support for individuals and supply

political leverage and volunteers for community institutions” (Putnam 2000: 312).

The decline in the levels of social capital may pose a serious danger to the quality of life

in local communities, but also – more generally - to the functioning of political and economic

institutions (Inglehart 1999; Knack and Keefer 1997; Putnam 1993b). Norms and networks

constituting social capital have been treated as a powerful exogenous factor enhancing

institutional performance: trust, reciprocity and the sense of solidarity among citizens reduce

the costs of policy and rule implementation, transform citizens’ preferences from particularistic

into collectivist, increase predictability and reliability of economic and political transactions

(Boix and Posner 1998).3 However, in this paper social capital is of interest not because of its

purported effect on institutional functioning, but because it is a part of a broader phenomenon

of social cohesion.

Components of social capital

Social capital consists of two main components: cognitive and structural. Attitudes of

trust and reciprocity in relations with fellow-citizens are stimulated by the interpersonal

3 It should be noted that there exists a growing body of literature and research arguing an

opposite, “top-down”, approach to social capital and institutional performance, where social

capital is endogenized and treated as a consequence rather than cause of institutional setting.

See Jackman and Miller 1998; Letki and Evans 2005; Muller and Seligson 1994; Rose-

Ackerman 2001.

6

interactions. It has become a convention to use membership in voluntary associations as a

measure of citizens’ participation and interaction that leads to the emergence of the attitudes of

interpersonal trust and reciprocity. However, while voluntary associations have long been

recognized to perform the functions of “schools of democracy”, and to contribute to building

citizens’ capacity to participate in politics (Almond and Verba 1963; Edwards and Foley

2001), their ability to generate trust and reciprocity has been questioned. Stolle has found that

the positive relationship between interpersonal trust and associational membership is the result

of self-selection: trusting people are more likely to join associations, and the length of

membership is unrelated to an individual’s levels of trust in others. Uslaner has argued that

trust is a “moral value” and as such is independent of civic or associational activism, but

depends on the patterns of childhood socialization instead. More recently, Li, Pickles and

Savage have found out that voluntary associations are highly selective in their membership

policies, they reinforce existing social divisions (along the class and gender lines) and

contribute to the emergence of exclusive group identities (Li et al. 2005; Stolle 1998, 2001;

Uslaner 1999). Therefore, it is necessary to complement associational involvement with other,

informal, types of interactions. In this present paper I turn to two more indicators of

interpersonal relations: informal neighbourhood sociability and individual voluntary help given

to or received from fellow neighbours.

Informal sociability, although it does not serve any particular function other than

gathering or socializing itself, creates an opportunity to adjust reciprocal obligations among

friends or neighbours as members of the same community. Informal sociability promotes open

communication, interest in others’ problems or points of view, stimulates mutual care, trust

and understanding (Misztal 2000). Through informal sociability individuals form their social

networks freely, without being restricted by organizational rules. Offering help to a fellow-

citizen is an explicit expression of the sense of commitment and care. It contributes to the

emergence of the attitudes of reciprocity and consideration. The importance of informal

relations has been reflected in the theoretical work as well as in the empirical research on the

causes and consequences of social capital.4 And, while formal engagement in voluntary

associations is an indicator that is easier to obtain, informal sociability seems in fact to be

conceptually closer to the idea of interpersonal networks and spontaneous interaction that

underline the production of social trust and reciprocity.5 Informal sociability is also a more

4 For theoretical discussions, see Bourdieu 1986; Coleman 1990; Lin 2001, for empirical

research see Lowndes 2000; Marschall and Stolle 2004; Putnam 2000.

5 It seems that there is a conflict between political scientists’ approach stressing organisational

involvement as a virtuous generator of social trust (Putnam 2000) and sociologists’ approach

7

democratic form of interaction than membership in formal groups and associations, which is

strongly determined by a socio-economic status. While the well-educated are more likely to be

association members, it is the lower classes that have stronger neighbourhood links (Li et al.

2005). The link between attitudes of trust and reciprocity on one hand, and informal sociability

on the other is well described by Putnam, who speaking of different forms of socializing says

that “like pennies dropped in a cookie jar, each of these encounters is a tiny investment in

social capital” (Putnam 2000: 93). Therefore, to reconstruct social capital most fully, in this

paper I use indicators of both formal (organizational) and informal (social or individual)

interactions.

To sum up, the sense of community, neighbourhood attachment, trust and care about

fellow-neighbours are the result of various types of interactions. Some of these interactions are

formal, for example in voluntary associations, and some other occur only within informal

networks of friendly or neighbourly sociability or assistance. Interactions stimulate perceptions

and attitudes, but the relationship between various types of interactions is “mutually

reinforcing”. For example, individuals who are active socially are more likely to offer help to

others, those who are involved in voluntary associations are more likely to have friends and

know their neighbours, thus they are more likely to socialize informally, etc. As Costa and

Kahn put it, “social capital refers to aspects of the network structure—such as social norms and

sanctions, mutual obligations, trust, and information transmission—that encourage

collaboration and coordination between friends and between strangers” (Costa and Kahn 2003:

103).

Diversity, neighbourhood status and social capital

The concerns about the crisis of social cohesion in Britain and beyond have been

answered by the search into the factors that may re-invigorate it as much as the factors that

erode it. The first stream of investigation focused on “the practical problems associated with

attempting to formulate a public policy of community cohesion on the assumption that

common principles and shared values can be founded in multiethnic, multi-faith and multi-

cultural societies” (McGhee 2003: 376), while the other followed the evidence from other

countries suggesting that more homogeneous communities have higher levels of social

focusing on generally defined networks as resources an individual can use to their

advancement (Bourdieu 1986; Coleman 1990; Lin 2001). This rift also largely overlaps with

the division between scholars who treat social capital as a community or individual resource.

For a recent discussion of the commonalities and differences in particular approaches to social

capital, see Li et al. 2005.

8

interactions, which in turn leads to more social capital (Alesina and Ferrara 2000; Costa and

Kahn 2003). Both threads lead therefore to the same issue: diversity.

Diversity

There exists a number of theoretical and empirical works claiming that diversity, and

racial diversity in particular, seriously undermines a sense of community and social cohesion.

Alesina and La Ferrara identify a very powerful negative relationship between racial diversity

and levels of interpersonal trust across American states. To explain this pattern they refer to the

“natural aversion to heterogeneity” (Alesina and Ferrara 2002: 225). In another paper they

posit that diversity has a negative effect on interactions among individuals: “In our model

individuals prefer to interact with others who are similar to themselves in terms of income,

race, or ethnicity”, and this finding is confirmed by other researchers (Alesina and Ferrara

2000: 850; Costa and Kahn 2003; Putnam 2003).6 Experimental research shows that inter-

racial exchanges involve significantly lower levels of honesty and reciprocity (Glaeser et al.

2000). The most popular interpretation of this pattern refers explicitly to the “preferences of

homogeneity” or, in other words, principle of homophily: contacts among similar people occur

more often than contacts among dissimilar people (McPherson et al. 2001). Therefore,

individuals living in heterogeneous communities are expected to interact less frequently, which

in turn leads to lower levels of interpersonal trust and reciprocity.

This approach to explaining the negative effect of racial diversity on intra-community

relations should be supplemented with the findings from the research on racial attitudes and

racial prejudice. Research on racial attitudes in the US focusing on the impact of context on

attitudes has shown that white respondents’ racial hostility increases together with the increase

in racial diversity of the area they live in (Glaser 1994; Huckfeldt and Kohfeld 1989; Taylor

1998). The explanations of the mechanism behind this relationship refer to the perceptions of

‘power threat’: dominant group is hostile towards minority groups over the economic and

social privileges they fear to lose.7 However, a different approach, stressing contact rather than

context, shows that any type of inter-racial contacts and residential integration significantly

reduce inter-group prejudice. Oliver and Wong found out that living in heterogeneous

6 It should be noted, however, that Costa and Khan seem “troubled” with their findings that

segregation promotes social capital. They conclude their paper saying that: “From society's

perspective, racial and ethnic equality and equality of access may be more important values

than achieving greater civic participation” (Costa and Kahn 2003: 109).

7 This line of argumentation is echoed in the British media and political discourse on the

eroding effect of diversity on social cohesion and welfare state. See Goodhart 2004, also The

Economist 2004.

9

neighbourhoods is correlated with positive attitudes towards out-groups, while “those who live

amongst more of their own racial group hold more negative views of out-groups and perceive

more competition from out-groups” (2003: 568). Bledsoe et al have found that blacks living in

racially mixed neighbourhoods feel less solidarity with other blacks than those who live in

more homogenous (i.e. more black) neighbourhoods (Bledsoe et al. 1995, see also Stein et al.

2000). The positive effect of interactions in the context of racial diversity on more general

attitudes, such as interpersonal trust, has been shown by Marschall and Stolle (2004).

The findings discussed above seem at first contradictory, although it is perfectly

plausible that both sides of the argument are right. Diversity may well have a negative effect

on individuals’ propensity to interact with fellow neighbours, however, once the interaction

takes place, its effect is positive: the attitudes of racial hostility and prejudice are overcome,

and an individual becomes more favourable towards other people in general. What is equally

important, and what few of the studies quoted above take into account, is the contextual

characteristic that is usually strongly correlated with racial diversity: low socio-economic

status of a community.

Neighbourhood status

Research on the community-level determinants of both trust and racial attitudes has

shown that the socio-economic status of a neighbourhood an individual is living in has a

crucial influence on the formation of their attitudes. When neighbourhood socio-economic

context (apart from just individual’s socio-economic status) is taken into account, it turns out

to be the main factor triggering negative attitudes and lack of trust in out-group members.

Oliver and Mandelberg have found out that physical and economic duress trigger “interracial

material competition”, which in turn generates negative attitudes towards of members of other

racial groups. Ross et al. identified neighbourhood economic disadvantage and disorder as the

main factors eroding interpersonal trust. Li et al. established that high socio-economic ward

status is among the important predictors of neighbourhood attitudes and participation.

Moreover, Ross et al. have identified an amplifying effect of neighbourhood disadvantage on

the effect of individual-level disadvantage: the combination of low individual-level status and

low neighbourhood status are particularly damaging for trust (Li et al. 2005; Oliver and

Mandelberg 2000; Ross et al. 2001, see also Marschall and Stolle 2004). High levels of

unemployment, crime and disorder generate the feelings of powerlessness, threat and

alienation among residents, which in turn lead to the low levels of neighbourhood attachment

and interactions:

10

“Low-status settings … expose residents to a daily dose of petty crime, concentrated physical

decay and social disorder … This exposure in turn leads to a constellation of negative

psychological states which are experienced by residents: feelings of anxiety and fear, alienation

from neighbours, lack of trust in others, and suspicion toward out-groups in general” (Oliver and

Mandelberg 2000: 576).

Summing up, deprivation and disorder damage the sense of community: they erode

formal and informal interactions, which in turn lowers interpersonal trust and sense of

belonging to a neighbourhood. Additionally, deprivation and disorder tend to be strongly

correlated with racial diversity, which is not sufficiently accounted for by most methods

customarily used to analyse determinants of community attitudes at the aggregate level. As a

result, they create the perception of diversity being the main causal factor undermining social

cohesion.

Reprise

The sections above have outlined the general theoretical background to the research into

the link between racial diversity and social capital. Here, I recapitulate the main points.

Neighbourhood social capital consists of two major components: attitudes towards fellow

neighbours and interactions with them. The first component invokes the sense of attachment to

the community, trust in neighbours, familiarity and care. The second component includes

formal and informal interactions, such as activism within groups and organizations, helping

others and informal sociability with friends and neighbours. Interactions generate positive

neighbourhood attitudes, but they also stimulate each other. For example, individuals with a

wide circle of friends are more likely to volunteer to help someone, while active group and

organization members are more likely to have a rich social life and vice versa. These

relationships are purported to hold at the individual level as well as community level. So,

neighbourhoods with a lively social life will also have strong networks of friendship and high

levels of neighbourly help. As a result, even individuals not directly involved in any of these

activities will have more positive opinions and attitudes about neighbours and neighbourhood.

Racial diversity erodes interactions and, as a result, changes people’s attitudes towards

their neighbourhood. As individuals prefer to socialize with those who are like themselves, in

diverse communities levels of interpersonal contact and interaction will be lower. Another

neighbourhood-level factor determining social capital is socio-economic status: high levels of

disorder and deprivation generate powerlessness and mistrust. They also limit individuals’

interactions: infrastructure (such as meeting venues) is scarce and staying in the public places

is not safe, both for individuals and their property. Low socio-economic status of a

neighbourhood tends to be highly correlated with racial diversity.

11

Figure 1 about here.

Figure 1 is a simplified schematic presentation of the model outlined above. Ovals

represent components of the model, and arrows illustrate the direction of causality of

relationships between them (with double-ended arrows representing a bi-variate relationship).

The relationships between particular components of social capital apply to the individual as

well as community level: neighbourhoods with rich organizational life are expected to be

supportive and socially active communities, where there is a high degree of mutual trust and

commitment. Neighbourhood social capital is simultaneously affected by neighbourhood-level

and individual-level characteristics.

Data and indicators

Although theoretical discussion of the causes and consequences of social capital refers

to “community” or “neighbourhood” level explanations, the empirical evidence is usually

based on countries or states/regions as units of analysis (Inglehart 1999; Newton 1999; Paxton

1999; Putnam 1993a). Notable exception are studies by Li et al. (2005), Marshall and Stolle

(2004) and Oliver and Wong (2003), who analyze the impact of contextual (neighbourhood or

ward level) characteristics on individuals’ attitudes towards others. This present paper follows

their approach in that it uses the actual neighbourhoods as one of the levels of analysis. When

investigating correlates of social capital it is crucial to look at the neighbourhood rather than

municipality, region, or country, because the postulated mechanisms explaining the generation

of social capital refer to the familiarity and bonds that can only be formed at a very local level.

Therefore, I employ data collected with the purpose of analyzing neighbourhood effects:

individuals interviewed were clustered within postcode sectors.

Data were collected between March and October 2001 in England and Wales as a part of

the Home Office Citizenship Survey. The Postcode Address File was used as a sampling frame

and “six addresses neighbouring each core sample address were identified by interviewers in

the field, using a strict set of rules… At each eligible sampled address an individual aged 16 or

over was randomly selected for an interview” (Attwood et al. 2003: 2). The main sample was

complemented with the ethnic minority boost sample, resulting in the total sample of 15093

individuals nested within 839 neighbourhoods. The average number of respondents per sample

is 18.09. The neighbourhood-level information on racial composition of neighbourhoods has

12

been obtained from the 2001 Census and the index of deprivation comes from the Office of the

Deputy Prime Minister.8

Social capital

In this paper I use a wide range of indicators that refer to various aspects of

neighbourhood social capital, such as attitudes towards fellow neighbours, formal and informal

interactions. These indicators, displayed in Table 1 below, focus on neighbourhood, which

rules out ambiguities most of the survey items used by other authors suffer from.9 The first two

questions (a and b) asked respondents whether they enjoy living in their neighbourhood and

whether people in this area look out for each other. 63.0% of respondents answered that they

definitely enjoy living in their area (the other options for both questions were to some extent

and no) and 40.4% stated that in their area neighbours definitely look out for each other. 58.8%

and 67.0% of respondents (items c and d) believe that, respectively, they know some or many

people in their neighbourhoods and that some or many people can be trusted (the other options

were few people and none). 32.4% of respondents believe that it is very likely or quite likely

that their wallet or purse, if lost in the neighbourhood, would have been returned with nothing

missing from it (item e).

Items f to h refer to informal interactions with friends and neighbours. Respondents were

asked to indicate on a seven-point scale (ranging from never to every day) how often they have

friends or neighbours round to their house, visit them, or go out together, to a pub, restaurant,

cinema etc. As many as 50.3% of respondents stated that they invite friends or neighbours at

least once a week or more often (several times a week or every day). A similar proportion of

respondents (45.1%) indicated that they visit friends and neighbours at least once a week.

Finally, 32.7% of respondents go out socially with a group of friends or neighbours once a

week or more often.

Table 1 about here.

8 Accessed at, respectively, http://www.statistics.gov.uk/census2001, http://www.odpm.gov.uk.

The datasets were merged by Giselle Baker and Patten Smith from BMRB.

9 The most popular indicators of social capital, interpersonal trust measured by means of a

question “Most people can be trusted”, and membership in associations, present some

interpretation problems, as it is not entirely clear how respondents define “most people” and

whether membership entails any actual interaction (in contrast with passive membership,

chequebook participation etc.)

13

Items i to l refer to the frequency of formal or informal volunteering and receiving help

from formal or informal volunteers in the last 12 months. The possible answers were at least

once a week, less than once a week but at least once a month, and less than once a month.

49.7% of respondents are involved in activities of a formal group, club or organization at least

once a month. Because the format of the question taps directly into active involvement in the

relatively recent period of time (within the last 12 months), this question discriminates against

pocket book participation or inactive membership, which would be very unlikely to generate

any neighbourhood-related opinions and attitudes. The next question asked about frequency of

volunteering to help organizations, groups or clubs, for example by raising money, doing

clerical work, campaigning, providing transport etc. 25.6% of respondents did at least one of

these things once a month or more often. Frequency of helping others as an individual (rather

than a member of an organized group) is captured by item k: 33.0% of respondents helped

friends or neighbours at least once a month by, for example, keeping in touch with someone

who has difficulty getting out and about, doing shopping, cooking or cleaning, baby-sitting,

decorating, giving advice, representing or providing transport. A lower proportion, 23.0%,

received unpaid help from others, either acting on behalf of a group or as individuals (item l).

Overall, these figures indicate high levels of trust, care and sense of connectedness in

British neighbourhoods. Britons are also very sociable and active in formal organizations.

When these indicators of neighbourhood social capital are subjected to factor analysis, they

form four dimensions: attitudes and opinions about neighbours and neighbourhood (items a to

e), informal sociability (f to h), formal volunteering (i to j) and informal help (k to l). The

results of exploratory factor analysis are displayed in Table A1 in the Appendix. The same

table contains information on intra-class correlations of indicators of social capital and the

magnitude of design effects, which show that a multi-level model is necessary to analyze

neighbourhood social capital adequately.10

Diversity

To measure ethnic diversity I follow the convention of discussing the issues of ethnicity

in terms of race (Saggar 2000). Previous research also focused predominantly on racial

composition of an area as an explanatory factor of attitudes and behaviour (Costa and Kahn

2003; Hero 2003; Marschall and Stolle 2004; Oliver and Mandelberg 2000). I use racial

fragmentation as a measure of ethnic diversity. It is computed according to the following

formula:

10 The “rule of thumb” is that the design effect of 2 or higher requires a multi-level model.

14

Racej = 1 - ∑ S2kj

(1)

where j stands for a neighbourhood area and k for the following racial groups: (a) White,

(b) Black and mixed Black, (c) Asian and mixed Asian, (d) Chinese and mixed Chinese, and

(e) other. Term Skj represents a proportion of a given group in the neighbourhood. The index

increases as heterogeneity increases: 0 means perfect homogeneity, and 1 means that every

person living in a neighbourhood belongs to a different racial group. As a result, it captures the

probability of two randomly chosen individuals within one neighbourhood being members of

different racial groups. The index ranges from 0 to .729, with an average of .297 and SD of

.244 (descriptive statistics can be found in Table A2 in the Appendix).

Neighbourhood socio-economic status

To measure the socio-economic status of a neighbourhood I use an Index of Multiple

Deprivation, covering the following six domains: income, health deprivation and disability,

employment, education, skills and training, housing and geographical access to services

(Department of the Environment 2000).11 The full range of this index is from 0 to 100, where

100 is the highest level of deprivation, but it varies across neighbourhoods present in the

sample from 2.19 to 77.58, with the mean at 33.74. As previous research suggests, there is a

strong positive relationship between low neighbourhood status and racial heterogeneity, as the

correlation between these two variables at the neighbourhood level is r = .554, at p <.001

(Sampson and Groves 1989; Sun et al. 2004).

Individual level determinants of social capital

While this paper focuses predominantly on the influence of contextual effects on social

capital in British neighbourhoods, it is necessary to account for individual-level determinants

of social capital. Education is one of the main correlates of trust and formal involvement and

social class is an important indicator of resources necessary for formal and informal

participation. Prior research on the determinants of social capital in Britain has identified class

and education level differences as the key determinants of various dimensions of social capital,

as groups with different types of resources tend to “specialise” in different types of social

capital (Hall 1999; Li et al. 2005). Age is usually increasing probability of individuals forming

positive attitudes towards and opinions about other people, but it may inhibit formal and

11 The index, as well as crime levels, are measured at the ward level, which is a higher level of

aggregation than the neighbourhood. This is not problematic for the analysis, as less than 16%

of neighbourhoods are not uniquely assigned to a ward.

k

15

informal engagement. On the other hand, the declining levels of formal and informal activism

as well as social trust are frequently linked to the generational change (Putnam 2000).

Therefore, I account for these factors at the individual level. The average age in the sample is

46 years. 16.6% of respondents have a degree, and 30.7% belong to the salariat - higher or

lower managerial and professional occupations (Brehm and Rahn 1997; Li et al. 2005;

Marschall and Stolle 2004; Uslaner 1999).

It would be natural to expect, especially in the light of earlier research on racial attitudes

and on social trust and participation, that one’s race or ethnicity is an important predictor of

attitudes towards fellow neighbours and formal and informal participation in the context of

British neighbourhood. However, while race is certainly important for explaining political

attitudes and behaviour in Britain (Heath and Saggar 1999; Saggar 2000), its predictive power

in the models tested in this paper was marginal, and including it significantly worsened the

model fit, therefore it has been excluded from the final model. The lack of relevance of race in

the UK can be explained by an entirely different history of racial minorities in the UK and US,

which shaped differently the sense of “linked fate” and inter-racial relations in both countries.12

Method

2001 Citizenship Survey addresses the issue of social capital through a range of

questions related to the various aspects of participation in a community, organizational

involvement, friends’ networks and, most importantly, opinions about community and

neighbourhood. The data, as mentioned above, have a hierarchical structure: individuals are

clustered within postcode sectors. To make use of such nested structure of the data, I will use

hierarchical linear models, which allow for partitioning variance and covariance components

between the levels (i.e. they allow for estimating the extent to which the variance in

individuals’ attitudes or behaviour is explained by the individual and neighbourhood-level

factors).13 They also allow for the simultaneous estimation of the impact of individual and

neighbourhood level characteristics on individuals’ attitudes and prevent an underestimation of

the standard errors of the coefficients, especially at the higher level, which might in turn lead

to the inference that effects are significant when they are not (Bryk and Raudenbush 1993;

Woodhouse et al. 1996).

12 Previous research has demonstrated some differences in social capital accumulation among

ethnic minorities in Britain (Li 2005). However, these results were not confirmed in this

present research.

13 This type of models is also often referred to as multilevel linear models, random-coefficient

regression models and covariance components models.

16

The multi-level model can be summarized by the following equation:

yin = β0inx0 + β1x1in… + βzxn (2)

Subscript in denotes that a given variable varies between individuals and

neighbourhoods, and subscript n denotes that a variable varies between neighbourhoods and is

constant for all respondents within a given neighbourhood. β0in is an intercept explained by the

formula

β0in = β0 + u0n + e0in (3)

where u0n is a level 2 residual (constant for all respondents within a given

neighbourhood, but varying between neighbourhoods), while e0in is a level 1 residual, varying

between persons and neighbourhoods. u0n and e0in are assumed to be uncorrelated.

Because of the implied relationships between particular variables, it is necessary to use a

path model. This means that a number of linear regressions will be estimated simultaneously at

both the individual and neighbourhood level. Finally, because a number of indicators will be

used to capture underlying concepts, such as informal sociability or neighbourhood socio-

economic status, it is necessary to create latent variables at both levels. Therefore, the method

used is a multilevel structural equation model, combining path analysis and factor analysis. A

model of this sort combines the advantages of hierarchical approach by accounting for the

design effect and latent variable approach by adjusting for the measurement error.14 I use

Mplus 3.12, software created with the purpose of fitting models with latent variables.15

Neighbourhood social capital and racial diversity: Testing the relationship

The main dependent variable of the tested model is neighbourhood social capital, which

is regressed at the individual level on age, education and social class, and at the neighbourhood

level on socio-economic status and racial diversity. Social capital consists of four dimensions,

which are latent variables with two to five observed indicators. The dimensions of social

capital are interrelated: informal sociability, organizational involvement and individual help

are correlated with each other, and all three of them influence attitudes towards and opinions

about neighbourhood. Additionally, there are correlations between some of the observed

indicators (for example, group involvement and going out), which is accounted for in the

model. At the neighbourhood level, socio-economic status is measured by means of an Index

of Multiple Deprivation. Since in this paper I hypothesize that racial diversity is negatively

tinted by a low socio-economic status, I account for the correlation between the two.

14 For the discussion of multilevel structural equation models, see Hox 2002.

15 Program Copyright © 1998-2004 Muthén & Muthén.

17

Modelling the relationship between racial diversity and community socio-economic status in

this way responds to the criticisms against treating race as an exogenous factor and ignoring its

social and economic aspects (Bedolla and Scola 2004; Smith 2004).

Table 2 about here.

Table 2 above displays coefficients from the multilevel structural equation model testing

the effect of individual and neighbourhood characteristics on social capital. Apart from

unstandardised estimates and their standard errors, it presents standardized coefficients, which

make it possible to compare the magnitude of particular effects across the model. Effects

missing a conventional .05 significance level have been italicized. Proportion of particular

variables’ variance explained is presented in the final column. The model has been estimated

simultaneously for the within and between levels.

Both ‘within-level’ and ‘between-level’ parts of the table begin with displaying factor

loadings of the observed variables that have been used to capture endogenous (Neighbourhood

Attitudes, Sociability, Organisational Involvement, Individual Help) latent variables. The first

factor loading for each latent variable is fixed to 1. These sections of the table are followed by

the estimates from the regressions of the components of social capital on individual and

neighbourhood level explanatory variables. The table also presents estimates for the bi-variate

correlations between the latent (e.g. Sociability and Organisational Involvement) or observable

(e.g. Trust with Know) variables.16

Indicators of model fit show that the model fits the data well: CFI is .956, which is

slightly higher than the conventional cut-off point for this index. Other indices also suggest a

good fit: SRMR for both within and between levels falls well below a recommended cut-off

point of .08, and RMSEA is also low at .023 (the recommended cut-off point for this index is

.06). The lower value of SRMR for the between-level (.026) than for the within-level (.049)

shows that the relationships specified across neighbourhoods fit the data somewhat better than

those specified for within the neighbourhoods. Most of the effects at both levels are highly

statistically significant, but those that miss the .05 significance level are italicized. Weights

have been used to account for the overrepresentation of minority respondents.

16 Correlations between some observed indicators of social capital (e.g., trusting and knowing

the neighbours) at the individual level were added because a model with no correlated error

terms between these items did not fit the data adequately.

18

Discussion of the findings

Dimensions of social capital

The main question of this paper is whether racial diversity undermines social capital in

British neighbourhoods or whether its seemingly destructive effect is a result of its empirical

association with low neighbourhood status. Before answering it I should, however, look at the

dimensions of social capital and links between them. Does informal socialising, joining

associations and helping people informally stimulate positive attitudes to one’s

neighbourhood? Are people who socialise also the ones who join groups and help others? Are

neighbourhoods with higher levels of formal and informal interactions among their inhabitants

also full of trust and reciprocity? The answer is “yes”, yet with some caveats. Behavioural

dimensions of social capital – sociability, organisational involvement and informal help – are

positively correlated at the individual level, and while similar relationships are present at the

neighbourhood level, there is one exception: the most sociable neighbourhoods are not the

ones with the highest levels of organisational participation (as this effect is statistically

insignificant). Moreover, individuals living in neighbourhoods where informal help is

frequently offered and received have more negative opinions about their local community. This

is a puzzling finding, yet it may be interpreted as showing that high levels of formal and

informal assistance in a neighbourhood indicate an additional degree of disadvantage, not

captured by the objective measure of deprivation. Thus, individuals living in communities with

strong help networks may be in fact living in particularly disadvantaged neighbourhoods,

which in turn would negatively affect their opinion about these communities.

At the individual level, sociability strongly stimulates positive attitudes towards fellow

neighbours (β = .259), while organisational involvement, despite being celebrated as one of the

key determinants of trust and reciprocity, has only a weak effect (β = .085). These results are

consistent with earlier findings, stressing the importance of informal networks for generating

trust and reciprocity (Li et al. 2005) and highlighting the ambiguities of the link between

formal involvement and trust (Stolle 1998). It also seems that people involved in the reciprocal

networks of help tend to be active both informally and in organisations (positive correlations of

Individual Help with Sociability and Organisational Involvement). However, giving help to

friends or neighbours and receiving it does not make individuals perceive their neighbourhood

more favourably (the effect of Individual Help on Neighbourhood Attitudes is statistically

insignificant). One possible interpretation of this finding refers to the fact that seeking or

offering help is likely to take place within the already existing networks (built, for example,

through informal sociability, which is suggested by a positive association between Help

Received and Friends), thus not affecting individuals’ attitudes towards their neighbourhood.

19

The difference between formal and informal networks as the determinants of

neighbourhood attitudes is likely to be reflecting the distinction between what Putnam has

called schmoozers and machers (Putnam 2000; see also Li et al. 2005). It seems that in Britain

these are the schmoozers’ spontaneous, informal contacts with fellow citizens rather than

machers’ organised associational activities that lie at the heart of social cohesion and social

solidarity. Furthermore, there are some additional positive correlations between indicators of

particular dimensions of social capital which reveal certain patterns of behaviour: people who

think they know others well, trust them (correlation between Trust and Know). Informal

sociability is a reciprocal phenomenon: if we invite friends over, they will invite us, and vice

versa (correlation of Visit with Friends). Group involvement stimulates going out (correlation

of Group Involvement and Going Out), and people tend to receive assistance from people they

socialise with (correlation of Help Received with Friends).

Individual level determinants of social capital

The effects of socio-economic characteristics on particular dimensions of social capital

vary. Age has a strong and very strong effect on all four dimensions, but their direction differs:

it is positive in the case of attitudes, and negative in the case of behaviour. Older people are far

more likely to have a positive image of their local community and enjoy living there. In fact,

age is the single strongest determinant of neighbourhood attitudes (β = .302). It is also the

strongest factor influencing sociability, but in this case its effect is negative (β = -.504). Older

age also restricts other forms of formal and informal involvement, for example in organisations

or self-help activities. In addition, although age is a positive determinant of neighbourhood

attitudes, its indirect effect (via sociability and organisational involvement) is negative (in the

case of sociability it is as strong as β = -.130).

Indicators of socio-economic status – education and social class – have positive effects

on all four dimensions of social capital, but in some cases their strength is negligible. So far,

all empirical investigations of trust and reciprocity have identified a very strong effect of social

position on trust and reciprocity, and attributed it to the greater sense of control and higher

degree of life satisfaction that are associated with a higher socio-economic position. However,

since in the present model the characteristics of the respondents’ immediate surroundings are

accounted for by means of the level of neighbourhood deprivation and racial diversity,

individual-level indicators of socio-economic position have only a very weak effect on

respondents’ attitudes towards their fellow neighbours. This finding suggests that unless

contextual characteristics are controlled for, individuals’ socio-economic position is likely to

act as a proxy for the quality of the context in which attitudes of trust, care and reciprocity are

20

formed. Individuals of a higher socio-economic status tend to live in nicer, cleaner and safer

neighbourhoods, and when this is accounted for, the sole effect of their individual social status

is very weak.

In the case of informal sociability, the socio-economic status of an individual is of an

even lower importance: education has no significant effect and the effect of class is very weak.

Their impact on giving or receiving help is slightly stronger, but still limited. At the same time,

both education and class significantly increase people’s propensity to participate in formal

organisations. The fact that socio-economic status is an important covariate of organisational

membership, especially in comparison with more informal activities, confirms earlier findings,

where formal group membership has been identified as being strongly dependent on material

and cultural resources, and thus exclusive and reinforcing the existing social and economic

divisions. As a result, the reality of organisational engagement seems to be rather far from the

celebrated ideal of an all-encompassing civic activism and civil society. In contrast, informal

sociability is much less resource dependent and thus a more democratic form of participation

in the community life (Li et al. 2003, 2005).

Neighbourhood-level determinants of social capital

Now, we can turn to the effects neighbourhood-level characteristics have on social

capital. This question is answered by the “between-level” part of the model. From the

regressions of four components of social capital on neighbourhood characteristics it is clear

that low neighbourhood status has a detrimental effect on all four of them, and all these effects

are highly statistically significant. Comparing the standardised effect of neighbourhood status

on particular dimensions of social capital it is clear that it very strongly undermines

neighbourhood attitudes (β = -.449), organisational involvement (β = -.319) and sociability (β

= -.297); in the case of these three dimensions it is the single strongest determinant, even

comparing its effect with individual-level predictors (such as age, education or social class).

Individual help is also negatively influenced by low neighbourhood status, but to a lesser

degree (β = -.131). It should also be noted that the effect of neighbourhood deprivation on

neighbourhood attitudes is additionally transmitted through sociability and organisational

involvement, and these effects are β = -.073 and β = -.071 respectively. The combined direct

and indirect effects of deprivation on neighbourhood attitudes is, therefore, as high as β = -

.593.

In contrast, racial diversity negatively influences only one aspect of social capital:

neighbourhood attitudes. This effect is strong (β = -.345) and highly statistically significant,

yet it is the only negative effect of racial diversity detected. Accounting for the relationship

21

between diversity and deprivation (the standardised coefficient for this association is .363)

instead of entering them as two independent effects allows us to see that while in more racially

diverse neighbourhoods respondents indeed have more negative attitudes towards fellow

neighbours, this is not because they socialise or interact with them less (since diversity has no

effect on the level of formal and informal interactions in the neighbourhood). Therefore, claims

about the “natural aversion to heterogeneity” leading to more limited levels of interaction in

diverse communities and, in turn, to less trust and solidarity have not found any empirical

support.

Other tests

In terms of variance across respondents, sociability is definitely best explained at 25.7%.

Individual-level relationships explain only 11.7% of the variance of social trust, and less than

6% of organisational involvement and individual help. At the aggregate (neighbourhood) level,

the model explains as much as 68.9% of variance of neighbourhood attitudes, and 11.9% of

levels of organisational involvement. Levels of individual help and sociability are much less

well explained, with R2 = 1.5% and 8.9%, respectively.

Some additional tests were performed to examine the possible effect of an interaction

between individual-level status and neighbourhood status on attitudes towards fellow

neighbours. Following findings by Ross et al. and Oliver and Wong that individuals with low

socio-economic status (low education above all) are particularly adversely affected by the

contextual characteristics, I constructed a measure of individual-level deprivation using a range

of indicators, such as low level of education, council tenancy, unemployment, lone parenting

(Oliver and Wong 2003; Ross et al. 2001). However, these indicators did not form a single

dimension. Individually, they did not form interactions with the neighbourhood status that

would have statistically significant effect on neighbourhood attitudes. Neither did indicators of

a respondent’s high socio-economic status. So, the low status of neighbourhood negatively

affects people’s opinions about their fellow neighbours and local community no matter their

individual socio-economic status.

Conclusion

The main findings of this paper can, therefore, be summarised as follows: social capital

is a multifaceted and multi-layered phenomenon, and its various dimensions have different

individual-level determinants. The dimensions of social capital have a multi-level structure, but

this is particularly true for the neighbourhood-related attitudes. Although most of the

relationships between dimensions of social capital are statistically significant and have

22

expected, i.e. positive, direction, there are some exceptions. Informal sociability strongly

stimulates positive attitudes towards fellow neighbours, and this relationship holds for both

individual and neighbourhood levels of analysis. Organisational involvement, despite being

celebrated as one of the key determinants of trust and reciprocity, has a strong effect on

neighbourhood attitudes only at the community level, while at the individual level this effect is

much weaker. Giving or receiving informal help from friends or neighbours does not make

individuals perceive their neighbourhood in a more positive way, and the neighbourhoods with

higher average levels of help given or received have actually lower levels of positive attitudes

towards neighbours. Neighbourhood-level analysis suggests also that communities “specialise”

in either formal involvement or informal sociability (there is no correlation between

communities’ levels of informal sociability and organisational involvement). These findings

largely conform with the model of social capital proposing that contacts – both formal and

informal – facilitate familiarity and stimulate trust (Putnam 2000), although in British

neighbourhoods schmoozing seems to be a more effective way of enhancing social cohesion

than maching. Importantly, schmoozing is more resistant to individual-level deprivation than

formal activism. It should also be stressed that it is usually only one of these two types of

interactions that is at play, even when the socio-economic context is controlled for. Therefore,

analysis and operationalisations focusing on only one dimension or only one level of analysis

are unlikely to yield satisfactory and informative results. They also would not do justice to the

complexity and multidimensionality of the theoretical construct of social capital.

The importance of neighbourhood features for explaining social capital, in comparison

with the rather limited significance of respondents’ individual status-related characteristics,

sheds more new light on the process of social capital formation: the quality of the context in

which interactions are taking place seems more important for generating trust and reciprocity

than individuals’ socio-economic position. Therefore, without accounting for the contextual

characteristics we are unlikely to fully understand the determinants of formal and informal

activism and social trust.

However, apart from allowing us to disentangle these methodological issues related to

the empirical operationalisation of the phenomenon of social capital, the evidence presented

above leads to more interesting and powerful conclusions. It reveals that there is only very

limited empirical confirmation for the argument that racial diversity erodes social cohesion and

destroys relations in local communities in Britain. The point of departure for this research was

a concern expressed by the politicians, commentators and regular members of the public alike,

that modern Britain is becoming “too diverse” to maintain social cohesion and sense of

solidarity necessary for harmonious and steady development. They stress that the focus on

23

heterogeneity should be abandoned for the sake of common values and ideals, and that only

then self-help, communication and cooperation within society will be possible. However, this

present research has shown that when the association between racial diversity and economic

deprivation is accounted for, there is no evidence for the eroding effect of racial diversity on

interactions within local communities. It has also demonstrated, that interactions improve

perceptions of a neighbourhood, no matter its economic status or racial composition, but these

interactions are far less frequent in poorer neighbourhoods. There is no deficiency of social

capital networks in diverse communities, but there is a shortage of them in the economically

disadvantaged ones.

At the same time, racial diversity does have a direct negative effect on the perceptions

of, and trust in, fellow neighbours. This is a puzzling finding: although people living in racially

diverse neighbourhoods do not interact less with their neighbours, they declare less trust in

them and less satisfaction from living in their neighbourhood. One possible explanation for this

puzzle could point to the effect of media, such as framing or priming of racial attitudes and

inter-racial relations. The evidence from research into media effects in the US suggests that the

presence of explicit information and implicit clues about racial relations significantly

influences attitudes towards racial diversity (Kellstedt 2000). In particular, the racial “coding”

of crime and welfare in the minds of citizens results in the perceptions of diversity as

undesirable, which is frequently exploited in political campaigns (Valentino et al. 2002).

However, whether media is to be blamed for magnifying the negative associations of racial

diversity in Britain, is a question for further research.

Evidence presented in this paper has important policy implications. The negative effect

of diversity on the quality of social and civic life in the local communities in Britain, that

seems to have become a basis for the government’s strategy how to approach diversity is

largely spurious. There exists a relationship between solidarity, diversity and poverty.

Solidarity is undermined by poverty, but the blame is placed on diversity, as a result of the fact

that diversity and poverty are strongly associated.

In recent years ethnic and racial diversity has been increasingly identified as a key factor

undermining social cohesion and social solidarity. The British government, in pursuit of a

model for managing multiculturalism, has established a new strategy to deal with ethnic and

racial differences based on the assumption that, since the serious cohesion crisis experienced

by local communities is linked to diversity, the solution is to overcome differences and

strengthen the inter-community interactions and relations (Forrest and Kearns 2001; McGhee

2003). Reverting to neighbourhood relations, self help and community spirit as a panacea for

the social problems offers “an attractive (and cheaper?) alternative for tackling social exclusion

24

and regeneration” (Forrest and Kearns 2001: 2139). As a result, policies aimed at strengthening

social cohesion and solidarity for the purpose of either preserving the welfare state or

enhancing national identity focus on limiting the importance of cultural diversity in social life,

at the same time marginalising the importance of social and economic assistance (McGhee

2003).

This present research has not found evidence to support the thesis about the breakdown

of social connectedness and social life in the diverse communities in Britain. It did find,

however, that community cohesion and neighbourliness cannot be created in the context of

economic inequality and deprivation. We know that scarcity is one of the key factors triggering

conflict along racial and class lines (Burns and Gimpel 2000; Oliver and Mandelberg 2000).

We also know that ethnic groups face intergenerational disadvantage in terms of health,

education, and professional achievement (Roscigno 1998). Finally, individual-level

deprivation, no matter one’s ethnic background, limits opportunities to participate in social life

and exercise civic rights (Li et al. 2003, 2005), and neighbourhood deprivation is associated

with poor living conditions and disorder (Ross et al. 2001; Sampson and Groves 1989). Each

of these dimensions of inequality and deprivation individually offers an important explanation

as to why poverty destroys social cohesion and solidarity, but in real life they tend to coexist

and their effects cumulate. At the same time, each of them can be alleviated, but that requires

government intervention and cannot be achieved solely by reverting to “community relations”.

Therefore, the efforts to revive social cohesion through programs focused on inter-

community relations are misplaced if they under-emphasise material deprivation,

intergenerational disadvantage, crime and low community socio-economic status. To maintain

social solidarity and community cohesion 21st Century Britain needs more social and economic

equality, rather than more cultural unity. Until the link between diversity and deprivation is

alleviated, British communities are likely to continue to face a crisis of solidarity and collective

identity.

25

Figure 1. Neighbourhood Social Capital

Organisational

Volunteering

Individual

Help

Informal

Sociability

Neighbourhood

Attitudes

Nieghbourhood

Status Racial

Diversity

Socio-Economic

Characteristics

NEIGHBOURHOOD-LEVEL FACTORS:

INDIVIDUAL-LEVEL FACTORS:

Table 1. Indicators of Neighbourhood Social Capital

%

No Yes, to some

extent Yes, definitely

a) Enjoying living in neighbourhood 7.9 29.1 63.0

b) Neighbours look out for each other 17.6 41.8 40.6

None Few Some Many

c) Knowing people in neighbourhood 5.5 35.7 29.8 29.0

d) Neighbours can be trusted 25.5 7.5 34.8 32.2

Not likely at all Not very likely Don't know Quite likely Very likely

e) Wallet returned 31.3 28.3 8.0 24.1 8.3

Never Less than once a

month

At least once a

month

At least once a

fortnight

At least once a

week

Several times a

week Every day

f) Friends or neighbours over 10.4 15.0 12.7 11.6 23.6 18.3 8.4

g) Visiting friends or neighbours 13.8 15.6 13.4 12.1 23.8 16.7 4.6

h) Going out with friends or neighbours 23.4 19.4 15.0 9.6 19.8 11.2 1.6

Not at all Less than once a

month

Less than once a

week, but at least

once a month

At least once a

week

i) Taking part in organizations’ activities 38.1 12.2 13.7 36.0

j) Helping organizations 63.4 11.0 10.2 15.4

k) Helping a friend or a neighbour 38.1 28.9 18.2 14.8

l) Receiving unpaid help from organisations or individuals 50.9 26.1 12.0 11.0

Note: N= 15093

Table 2. Social Capital and Neighbourhood Context, N1=15093, N2=839

WITHIN-LEVEL b S.E. β R2

Neighbourhood Attitudes:

Enjoy 1.000 .000 .494 .244

Know 1.222 .068 .398 .159

Look Out 2.138 .078 .622 .387

Trust 2.270 .084 .583 .339

Wallet 2.134 .088 .500 .250

Sociability:

Friends 1.000 .000 .567 .322

Visit 1.258 .029 .717 .515

Go Out .974 .035 .566 .320

Organizational Involvement:

Group Involvement 1.000 .000 .741 .549

Group Help .948 .030 .775 .600

Individual Help:

Informal Help 1.000 .000 .664 .441

Help Received .706 .034 .495 .245

Regression of Neighbourhood Attitudes on .117

Sociability .073 .009 .259

Organizational Involvement .026 .006 .085

Individual Help .011 .012 .026

Age .005 .000 .302

Education (Degree) .030 .011 .039

Class (Salariat)

.044 .009 .071

Regression of Sociability on .257

Age -.027 .001 -.504


Class (Salariat)

.062 .030 .028

Regression of Organizational Involvement on .054

Age -.003 .001 -.071


Class (Salariat)

.253 .028 .128

Regression of Individual Help on .041

Age -.005 .001 -.142


Class (Salariat)

.126 .023 .087

Correlation of Sociability with Organizational Involvement .235 .016 .256

Correlation of Sociability with Individual Help .285 .015 .426

Correlation of Individual Help with Organizational Involvement .249 .012 .409

Correlation of Group Involvement with Go Out .246 .020 .116

Correlation of Trust with Know .148 .012 .153

Correlation of Visit with Friends .871 .045 .278

Correlation of Help Received with Friends .136 .015 .081

Note: Effects that are not statistically significant at .05 or below are italicized.

28

BETWEEN-LEVEL b S.E. β R2

Neighbourhood Attitudes:

Enjoy 1.000 .000 .786 .617

Know .880 .102 .517 .267

Look Out 1.576 .109 .767 .589

Trust 2.107 .137 .799 .639

Wallet 3.506 .203 .908 .824

Sociability:

Friends 1.000 .000 .744 .554

Visit 1.261 .155 .955 .911

Go Out .796 .132 .525 .275

Organizational Involvement:

Group Involvement 1.000 .000 .728 .530

Group Help .734 .075 .825 .681

Individual Help:

Informal Help 1.000 .000 .973 .947

Help Received .727 .091 .736 .542

Regression of Neighbourhood Attitudes on .689

Sociability .153 .038 .246

Organizational Involvement .119 .046 .222

Individual Help -.127 .043 -.218

Neighbourhood Status -.005 .001 -.449

Racial Diversity -.324 .034 -.345

Regression of Sociability on .089



Regression of Organizational Involvement on .119



Regression of Individual Help on .015


Racial Diversity .073 .088 .046

Correlation of Racial Diversity with Neighbourhood Status .987 .102 .363

Correlation of Sociability with Organizational Involvement .006 .007 .067

Correlation of Sociability with Individual Help .012 .007 .162

Correlation of Individual Help with Organizational Involvement .048 .008 .548

TESTS OF MODEL FIT: Chi2 = 1216.370 with 132 df; CFI = .956;

RMSEA = .023; SRMR Within = .049, SRMR Between = .026.

Note: Effects that are not statistically significant at .05 or below are italicized.

Appendix

A1. Dimensions of Social Capital: Standardized Factor Scores

I. II. III. IV. Intra-

class

Design

Effect

a) Enjoying living in neighbourhood .520 0.123 3.1

b) Neighbours look out for each other .630 0.115 3.0

c) Knowing people in neighbourhood .457 0.101 2.7

d) Neighbours can be trusted .666 0.146 3.5

e) Wallet returned .560 0.221 4.8

f) Friends or neighbours over .741 0.049 1.8

g) Visiting friends or neighbours .882 0.046 1.8

h) Going out with friends or neighbours .434 0.064 2.1

i) Taking part in organizations’ activities .985 0.046 1.8

j) Helping organizations .544 0.112 2.9

k) Helping a friend or a neighbour .729 0.058 2.0

l) Receiving unpaid help .439 0.080 2.4

Eigenvalue 2.023 2.840 1.384 1.038

Note: Exploratory factor analysis with Varimax rotation. N=15093, RMSA = .026.

Design effect = 1+(s-1)r, where s = average cluster size and r = intra-class correlation

30

Table A2. Exogenous variables: Descriptives

Min Max Mean SD

Neighbourhood Level, N = 839

Racial Diversity .000 .729 .297 .244

Deprivation 2.19 77.58 33.74 19.02

Individual Level, N = 15093

Age 16.00 98.00 46.39 18.70

Education (1 = Degree, 0 = else) .00 1.00 .17 .37

Class (1 = Salariat, 0 = else) .00 1.00 .31 .46

31

References

'The Kindness of Strangers. A Report on Multiculturalism', The Economist, 28 February 2004, 3-4

Alesina, A. and E. L. Ferrara, (2000), 'Participation in Heterogeneous Communities'; Quarterly

Journal of Economics, 847-904.

———, (2002), 'Who trusts others?' Journal of Public Economics, 85, 207-34.

Almond, G. A. and S. Verba, (1963), The Civic Culture: Political Attitudes and Democracy in Five

Nations Princeton: Princeton University Press.

Attwood, C., G. Singh, D. Prime and R. Creasey, (2003), 2001 Home Office Citizenship Survey:

People, Families and Communities London: Home Office Research, Development and

Statistics Directorate.

Baldwin, T. and G. Rozenberg, 'Britain "must scrap multiculturalism"', The Times, 3 April 2004

Bedolla, L. G. and B. Scola, (2004), Race, Social Relations and the Study of Social Capital, paper

presented at the American Political Science Association Meeting, Chicago, IL., September

1-4.

Bledsoe, T., S. Welch, L. Sigelman and M. Combs, (1995), 'Residential Context and Racial

Solidarity among African Americans'; American Journal of Political Science, 39, 434-58.

Boix, C. and D. N. Posner, (1998), 'Social Capital: Explaining Its Origins and Effects on

Governmental Performance'; British Journal of Political Science, 29, 686-93.

Bourdieu, P., (1986), 'The forms of capital', in Westport, J., ed., Handbook of Theory and Research

for the Sociology of Education, New York: Greenwood.

Brehm, J. and W. Rahn, (1997), 'Individual-Level Evidence for the Causes and Consequences of

Social Capital'; American Journal of Political Science, 41, 999-1023.

Bryk, A. S. and S. W. Raudenbush, (1993), Hierarchical Linear Models: Applications and Data

Analysis Methods Newbury Park: Sage.

Burns, P. and J. G. Gimpel, (2000), 'Economic Insecurity, Prejudicial Stereotypes, and Public

Opinion on Immigration Policy'; Political Science Quarterly, 115, 201-25.

Cantle, T., Community Cohesion: A Report of the Independent Review (London: Home Office,

2001)

Cheong, P. H., R. Edwards, H. Goulbourne and J. Solomos, (2005), Immigration, Social Cohesion

and Social Capital: A Critical Review, paper presented at the Whiter Social Capital? Past,

Present and Future, London, UK, 6-7 April.

Coleman, J. S., (1990), Foundations of Social Theory Cambridge, Mass., London: Belknap Press

of Harvard University Press.

Costa, D. L. and M. E. Kahn, (2003), 'Civic Engagement and Community Heterogeneity: An

Economist's Perspective'; Perspectives on Politics, 1, 103-11.

Denham, J., Building Cohesive Communities. A Report of the Ministerial Group on Public Order

and Community Cohesion (London: Home Office, 2001)

32

Department of the Environment, T. a. t. R., Regeneration Research Summary. Indices of

Deprivation 2000 (London: Department of the Environment, Transport and the Regions,

2000)

Devine, F., M. Savage, R. Crompton and J. Scott, (2003), Rethinking Class: Cultures, Identities

and Lifestyles Basingstoke: Palgrave.

Edwards, B. and M. W. Foley, (2001), 'Civil society and social capital: A primer', in Edwards, B.,

M. Foley and M. Diani, ed., Beyond Tocqueville. Civil Society and the Social Capital

Debate in Comparative Perspective, Hanover: Tufts University.

Erickson, R. and J. Goldthorpe, (1992), The Constant Flux: A Study of Class Mobility in Industrial

Societies Oxford: Clarendon Press.

Evans, G., ed., (1999), The End of Class Politics? Class Voting in Comparative Context, Oxford:

Oxford University Press.

Forrest, R. and A. Kearns, (2001), 'Social Cohesion, Social Capital and the Neighbourhood'; Urban

Studies, 38, 2125-43.

Fukuyama, F., (1995), Trust. The Social Virtues and the Creation of Prosperity London: Hamish

Hamilton.

Glaeser, E. L., D. I. Laibson, J. A. Scheinkman and C. L. Soutter, (2000), 'Measuring Trust'; The

Quarterly Journal of Economics, 115, 811-46.

Glaser, J., (1994), 'Back to the Black Belt: Racial Environment and White Racial Attitudes in the

South'; Journal of Politics, 56, 21-41.

Goodhart, D., 'Too Diverse?' Prospect, February 2004

Hall, P. A., (1999), 'Social Capital in Britain'; British Journal of Political Science, 29, 417-61.

Heath, A. F. and S. Saggar, (1999), 'Race: Towards a Multicultural Electorate?' in Evans, G. and P.

Norris, ed., Critical Elections: British Parties and Voters in Long-Term Perspective,

London: Sage.

Hero, R. E., (2003), 'Social Capital and Racial Inequality in America'; Perspectives on Politics, 1,

113-22.

Hox, J., (2002), Multilevel Analysis. Techniquees and Applications Mahwah, New Jersey:

Lawrence Erlbaum Associates, Publishers.

Huckfeldt, R. and C. Kohfeld, (1989), Race and the Decline of Class in American Politics Urbana:

University of Illinois Press.

Inglehart, R., (1990), Culture Shift in Advanced Industrial Society Princeton: Princeton University

Press.

———, (1999), 'Trust, well-being and democracy', in Warren, M. E., ed., Democracy and Trust,

Cambridge: Cambridge University Press.

Jackman, R. W. and R. A. Miller, (1998), 'Social Capital and Politics'; Annual Review of Political

Science, 1, 47-73.

Kellstedt, P. M., (2000), 'Media Framing and the Dynamics of Racial Policy Preferences';

American Journal of Political Science, 44, 245-60.

33

Knack, S. and P. Keefer, (1997), 'Does Social Capital Have an Economic Payoff? A Cross-

Country Investigation'; The Quarterly Journal of Economics, 62, 1251-88.

Letki, N. and G. Evans, (2005), 'Endogenizing Social Trust: Democratisation in East-Central

Europe'; British Journal of Political Science, 35, 515-29.

Li, Y., (2005), Social Capital, ethnicity and the labour market, paper presented at the International

Conference on Engaging Communities, Brisbane, 14-17 August.

Li, Y., A. Pickles and M. Savage, (2003), 'Social Capital and Social Exclusion in England and

Wales (1972-1999)'; British Journal of Sociology, 54, 497-526.

———, (2005), 'Social Capital and Social Trust in Britain the late 1990s'; European Sociological

Review, 21, 109-23.

Lin, N., (2001), Social capital: A theory of social structure and action Cambridge, UK and New

York: Cambridge University Press.

Lowndes, V., (2000), 'Women and social capital'; British Journal of Political Science, 30, 533–37.

Marschall, M. J. and D. Stolle, (2004), 'Race in the City: Neighborhood Context and the

Development of Generalized Trust'; Political Behavior, 26, 125-53.

McGhee, D., (2003), 'Moving to our common ground. A critical examination of community

cohesion discourse in twenty-first century Britain'; The Sociological Review, 51, 376-404.

McPherson, M., L. Smith-Lovin and J. M. Cook, (2001), 'Birds of a Feather: Homophily in Social

Networks'; Annual Review of Sociology, 27, 415-44.

Misztal, B., (2000), Informality: Social Theory and Contemporary Practice London and New

York: Routledge.

Muller, E. N. and M. A. Seligson, (1994), 'Civic Culture and Democracy: The Question of Causal

Relationships'; American Political Science Review, 88, 635-52.

Newton, K., (1999), 'Social Capital and Democracy in Modern Europe', in van Deth, J. W., M.

Maraffi, K. Newton and P. F. Whiteley, ed., Social Capital and European Democracy,

London: Routledge.

Oliver, J. E. and T. Mandelberg, (2000), 'Reconsidering the Environmental Determinants of White

Racial Attitudes'; American Journal of Political Science, 44, 574-89.

Oliver, J. E. and J. Wong, (2003), 'Intergroup Prejudice in Multiethnic Settings'; American Journal

of Political Science, 47, 567-82.

Ouseley, H., 'Forget this phoney debate, it's time to confront racism', Guardian, Date 2004, Issue.

Pakulski, J. and M. Waters, (1995), The Death of Class London: Sage.

Paxton, P., (1999), 'Is Social Capital Declining in the United States? A Multiple Indicator

Assessment'; American Journal of Sociology, 105, 88-127.

Portney, K. E. and J. M. Berry, (2001), 'Mobilizing Minority Communities. Social Capital and

Participation in Urban Neighborhoods', in Edwards, B., M. W. Foley and M. Diani, ed.,

Beyond Tocqueville. Civil Society and the Social Capital Debate in Comparative

Perspective, Hanover: Tufts University.

Putnam, R. D., (1993a), Making Democracy Work: Civic Traditions in Modern Italy Princeton:

Princeton University Press.

34

———, (1993b), 'The Prosperous Community. Social Capital and Public Life'; American

Prospect, 4, 35-42.

———, (1996), 'Who Killed Civic America?' Prospect, 7, 66-72.

———, (2000), Bowling Alone. The Collapse and Revival of American Community New York:

Simon and Schuster.

———, (2003), Ethnic Diversity and Social Capital, paper presented at the Ethnic Diversity and

Social Capital, ESRC Families and Social Capital Research Group Seminar, British

Academy, London, 24 June.

Roscigno, V. J., (1998), 'Race and the Reproduction of Educational Disadvantage'; Social Forces,

76, 1033-61.

Rose-Ackerman, S., (2001), 'Trust and Honesty in Post-Socialist Societies'; KYKLOS, 54, 415-44.

Ross, C. E., J. Mirowsky and S. Pribesh, (2001), 'Powerlessness and the Amplification of Threat:

Neighborhood Disadvantage, Disorder, and Mistrust'; American Sociological Review, 66,

568-91.

Saggar, S., (2000), Race and representation Manchester and New York: Manchester University

Press.

Sampson, R. and W. B. Groves, (1989), 'Community Structure and Crime: Testing Social

Disorganisation Theory'; American Journal of Sociology, 94, 774-802.

Sampson, R., S. W. Raudenbush and F. Earls, (1997), 'Neighbourhoods and Violent Crime: A

Multilevel Study of Collective Efficacy'; Science, 277, 918-24.

Smith, R. M., (2004), 'The Puzzling Place of Race in American Political Science'; PS: Political

Science and Politics, 37, 41-45.

Stein, R., S. Post and A. L. Rinden, (2000), 'Reconciling context and contact effects on racial

attitudes'; Political Research Quarterly, 53, 285-303.

Stolle, D., (1998), 'Bowling Together, Bowling Alone: The Development of Generalised Trust in

Voluntary Associations'; Political Psychology, 19, 497-525.

———, (2001), 'Clubs and Congregations: The Benefits of Joining an Association', in Cook, K. S.,

ed., Trust in Society, New York: Russell Sage Foundation.

Sun, I. Y., R. Triplett and R. R. Gainey, (2004), 'Neighbourhood Characteristics and Crime: A Test

of Sampson and Groves' Model of Social Disorganization'; Western Criminology Review,

5, 1-16.

Taylor, M. C., (1998), 'Local Racial/Ethnic Proportions and White Attitudes: Numbers Count';

American Sociological Review, 63, 56-78.

Uslaner, E. M., (1999), The Moral Foundations of Trust New York: Cambridge University Press.

Valentino, N. A., V. L. Hutchings and I. K. White, (2002), 'Cues that Matter: How Political Ads

Prime Racial Attitudes During Campaigns'; American Political Science Review, 96, 75-90.

Woodhouse, G., J. Rabash, H. Goldstein and M. Yang, (1996), 'Introduction to multilevel

modelling', in Woodhouse, G., ed., A guide for users of MLWin, London: Institute of

Education.

social capital and diversity final for PS that diversity, and racial heterogeneity in particular, is...

Documents

Transcript of social capital and diversity final for PS that diversity, and racial heterogeneity in particular, is...